update plot batch function

embedding_net/data_loader.py

@@ -44,7 +44,7 @@ def _load_images_paths(self):
         random.seed(4)
 
         for d in self.data_subsets:
-            print('=======================================')
+            print('{:5} ======================================'.format(d))
             self.images_paths[d] = []
             self.images_labels[d] = []
             for root, dirs, files in os.walk(self.dataset_path+d):
@@ -61,7 +61,6 @@ def _load_images_paths(self):
                 if curr_class in skip_list:
                     continue
 
-                print('Class {:11}: total number of files {}'.format(curr_class, n_obj))
                 idx_list = list(range(n_obj))
                 random.shuffle(idx_list)
                 count = 0
@@ -75,7 +74,7 @@ def _load_images_paths(self):
                 n_files_selected+=count
                 if d == 'train':
                     n_classes_selected += 1
-                print('Class {:11}: selected number of files {}'.format(curr_class, count))
+                print('Class {:11}: total number of files {:6}, selected {:6}'.format(curr_class, n_obj, count))
         print('Total number of files in dataset: {}'.format(n_files_dataset))
         print('Number of selected files: {}'.format(n_files_selected))
         print('Number of selected classes: {}'.format(n_classes_selected))
@@ -260,7 +259,7 @@ def get_batch_triplets_mining(self,
 
         all_embeddings_list = []
         all_images_list = []
-        # with_aug = s == 'train' and self.augmentations
+
         with_aug = self.augmentations
         for idx, cl_img_idxs in enumerate(selected_images):
             images = self._get_images_set(
@@ -348,6 +347,24 @@ def get_image(self, img_path):
                 img, (self.input_shape[0], self.input_shape[1]))
         return img
 
+
+    def plot_batch_simple(self, data, targets):
+        num_imgs = data[0].shape[0]
+        img_h = data[0].shape[1]
+        img_w = data[0].shape[2]
+        full_img = np.zeros((img_h,num_imgs*img_w,3), dtype=np.uint8)
+        indxs = np.argmax(targets, axis=1)
+        class_names = [self.classes['train'][i] for i in indxs]
+
+        for i in range(num_imgs):
+            full_img[:,i*img_w:(i+1)*img_w,:] = data[0][i,:,:,:]
+            cv2.putText(full_img, class_names[i], (img_w*i + 10, 20), cv2.FONT_HERSHEY_SIMPLEX,  
+                        0.5, (0, 255, 0), 1, cv2.LINE_AA)
+        plt.figure(figsize = (20,2))
+        plt.imshow(full_img)
+        plt.show()
+
+
     def plot_batch(self, data, targets):
         num_imgs = data[0].shape[0]
         it_val = len(data)